collect graph data

Author: bafgreat

README.md

@@ -60,13 +60,13 @@ It extracts **organic ligands**, **space group information**, and computes the *
 Quickly run command on any cif file
 
 ```bash
-fairmofsyncondition_syncon my_mof.cif
+  fairmofsyncondition_syncon my_mof.cif
 ```
 
 Or run and provide and outfile
 
 ```bash
-fairmofsyncondition_syncon my_mof.cif -o my_mof_report.txt
+  fairmofsyncondition_syncon my_mof.cif -o my_mof_report.txt
 ```
 
 `iupac2cheminfor`
@@ -75,13 +75,13 @@ as inchikey and smile strings directly from iupac names or common names. This ca
 achieved using `iupac2cheminfor` CLI as follows:
 
 ```bash
-iupac2cheminfor 'water'
+  iupac2cheminfor 'water'
 ```
 
 or
 
 ```bash
-iupac2cheminfor -n 'water' -o filename
+  iupac2cheminfor -n 'water' -o filename
 ```
 
 The out will be written by default to cheminfor.csv if no output is provided
@@ -93,15 +93,23 @@ Another useful tool is directly convert a `smile` or and `inchikey` their iupac
 To achieve this simply run the following commandline tool
 
 ```bash
-cheminfo2iupac -n 'O' -o filename
+  cheminfo2iupac -n 'O' -o filename
 ```
 
 `struct2iupac`
 In other cases one may one to directly extract the iupac name and cheminformatic identifier of a chemical structure.
 The quickest way to do this is by running the following commands.
 
 ```bash
-struct2iupac XOWJUR.xyz
+  struct2iupac XOWJUR.xyz
+```
+
+`pg_graph from cif file or folder`
+
+One can reliably create an `lmdb` dataset using the following code.
+
+```bash
+  pggraph_from_cifs -i ./CIFs/ -o mof_db.lmdb
 ```
 
 ## Training
@@ -123,7 +131,6 @@ command line arguments.
 find_bde_parameters -h
 ```
 
-
 ## Machine Learning Folder
 
 The folder **`machine_learning/`** contains the code and Jupyter notebooks to predict the **metal salt of a given MOF** using **Graph Neural Networks (GNNs)**.
@@ -143,8 +150,6 @@ The folder **`machine_learning/`** contains the code and Jupyter notebooks to pr
 **Note**:
 If you only want to **test the model** with pre-trained weights, you can simply **skip step 3** (training).
 
-
-
 ## Documentation
 
 Full documentation can be found [docs](https://bafgreat.github.io/fairmofsyncondition/).

fairmofsyncondition/call_model/run_model.py

@@ -12,7 +12,6 @@
 from pymatgen.symmetry.analyzer import SpacegroupAnalyzer
 from mofstructure.filetyper import load_iupac_names
 from fairmofsyncondition.read_write import cheminfo2iupac, coords_library, filetyper
-from fairmofsyncondition.read_write import filetyper, coords_library
 from fairmofsyncondition.call_model.utils_model import get_models, ensemble_predictions, get_model_energy, get_energy_prediction
 import warnings
 from openbabel import openbabel
@@ -325,7 +324,7 @@ def compile_data(self):
                 for solvent_data in solvents[:3]:
                     solvent, prob = solvent_data
                     data.append(f"{solvent:<40} {prob:>15} %\n")
-                    
+
         data.append("=" * 80 + "\n")
         data.append("\n")
         data.append("Report generated by fairmofsyncondition\n")

fairmofsyncondition/crystal/__init__.py

@@ -67,10 +67,10 @@ def find_diffraction_pattern(self):
         """
         if self.diffraction_type == "PXRD":
             xrd = XRDCalculator(wavelength=self.wavelength)
-            pattern = xrd.get_pattern(self.structure)
+            pattern = xrd.get_pattern(self.structure, two_theta_range=(0, 100))
         elif self.diffraction_type == "ND":
-            xrd = NDCalculator(wavelength=self.wavelength)
-            pattern = xrd.get_pattern(self.structure)
+            nd = NDCalculator(wavelength=self.wavelength, two_theta_range=(0, 100))
+            pattern = nd.get_pattern(self.structure)
         else:
             raise ValueError("""Invalid diffraction type.
                              Choose between PXRD and ND""")
@@ -81,8 +81,8 @@ def get_pattern(self):
         """
         """
         pattern = self.find_diffraction_pattern()
-        two_theta = np.array(pattern.x)
-        intensity = np.array(pattern.y)
+        two_theta = np.array(pattern.x, dtype=np.float32)
+        intensity = np.array(pattern.y, dtype=np.float32)
         return two_theta, intensity